How long will you live? This AI algorithm can predict when you will die with 78% accuracy (WATCH)
In a groundbreaking initiative, researchers in Denmark are delving into the intricate tapestry of human existence with the aid of artificial intelligence (AI) and vast datasets.
In a groundbreaking initiative, researchers in Denmark are delving into the intricate tapestry of human existence with the aid of artificial intelligence (AI) and vast datasets. Their project, known as Life2vec, aims to forecast the various stages of an individual's life journey, from infancy to the twilight years, shedding light on the possibilities and perils of AI technology in predicting human destinies.
Led by Sune Lehmann, a professor at the Technical University of Denmark (DTU), the team behind Life2vec is leveraging advanced deep-learning algorithms and data from millions of people to uncover patterns and relationships that shape human lives. Published recently in the journal Nature Computational Science, their study showcases the potential of AI to anticipate a wide array of health and social "life-events."
Life2vec operates on a framework similar to language-processing algorithms like ChatGPT but focuses on analyzing variables such as birth, education, social benefits, and work schedules. By scrutinizing sequences of life events, the algorithm endeavors to predict outcomes ranging from health indicators like fertility and obesity to socioeconomic factors like financial success.
The foundation of Life2vec lies in anonymized data sourced from approximately six million Danes, meticulously collected by the official Statistics Denmark agency. Through the analysis of event sequences, the algorithm demonstrates remarkable accuracy in predicting life outcomes, including mortality and relocation, with rates as high as 78% and 73%, respectively.
However, amidst the buzz surrounding Life2vec, concerns have emerged regarding privacy and the potential misuse of such predictive technology. Dubbed by some as a "death calculator," the project has attracted fraudulent entities offering life expectancy predictions in exchange for personal data, raising ethical red flags. Lehmann and his team emphasize that Life2vec remains strictly for research purposes and is not accessible to the wider public or commercial use.
Furthermore, the researchers aim to provide a public counterpoint to the opaque algorithms utilized by tech giants, highlighting the need for transparency and ethical considerations in AI development. Pernille Tranberg, a Danish data ethics expert, warns of the discriminatory implications of similar algorithms already deployed in industries such as insurance and healthcare.
Despite its promising capabilities, Life2vec is still in the nascent stages and not yet ready for practical application beyond the research realm. Lehmann and his colleagues aspire to delve deeper into long-term outcomes and the influence of social connections on life trajectories, further elucidating the complex interplay between data, AI, and human lives.